A Data Management Layer for Visual Information Retrieval

By Horst Eidenberger and Roman Divotkey

Abstract

This case study describes the data management layer of the VizIR visual information retrieval project. VizIR is an open source framework of software tools for visual retrieval research. In content-based multimedia retrieval media objects are described by high-dimensional feature vectors. These feature vectors have to be stored in an efficient way in order to accelerate the retrieval process. VizIR database management is based on object-oriented persistence management. The database interface has a three tier architecture: a pattern-based persistence system hides the underlying database, an object-relational mapping system maps classes to entities and a relational database provides state-of-the-art database features (transactions, integrity, recovery, etc.). The described database management prototype can be downloaded from the VizIR project website.

Reference

H. Eidenberger, R. Divotkey: "A Data Management Layer for Visual Information Retrieval"; Talk: ACM Multimedia Data Mining Workshop, Seattle, USA; 08-23-2004 - 08-26-2004; in: "ACM Multimedia Data Mining Workshop", (2004).

BibTeX

@inproceedings{TUW-138548,     author = {Eidenberger, Horst and Divotkey, Roman},     title = {A Data Management Layer for Visual Information Retrieval},     booktitle = {ACM Multimedia Data Mining Workshop},     year = {2004},     url = {http://publik.tuwien.ac.at/files/PubDat_138548.pdf},     keywords = {VizIR, Content-based Visual Information Retrieval, Video Retrieval, Image Retrieval, Object-oriented Database Design, Database Management, Persistence Management, High-dimensional Indexing, Multimedia Databases},     note = {talk: ACM Multimedia Data Mining Workshop, Seattle, USA; 2004-08-23 -- 2004-08-26} } Click into the text area and press Ctrl+A/Ctrl+C or ⌘+A/⌘+C to copy the BibTeX into your clipboard… or download the BibTeX.